This project seeks to identify how motor impairments in stroke survivors contribute to mobility deficits through the use of behavioral observations and computational models. Ultimately, this knowledge could be used to design more effective interventions to improve walking ability, increase functional independence, and reduce fall risk in individuals post-stroke. After a stroke, walking ability is affected by motor control deficits which are often characterized by hemiparesis, or weakness on one side of the body. During walking, hemiparesis is often associated with a visible limp that is due to differences in step lengths and stance times between the right and left sides of the body. Reducing these right-left asymmetries are a common objective of rehabilitation, and recently researchers have developed a number of approaches designed to reduce asymmetries such as acoustic pacing, unilateral step training, and split-belt treadmill training. Despite the recent clinical focus on reducing asymmetry, the potential functional benefits of improving symmetry have yet to be established. This research will address this gap by answering two fundamental questions: 1) Do improvements in symmetry lead to functional benefits such as a more efficient walking pattern or improvements in balance? 2) If stroke survivors retain the capacity to walk more symmetrically why do they choose to do otherwise? Although improvements in symmetry can no doubt reduce the potential stigma of walking with a limp, it is possible that a symmetric walking pattern could be less efficient or put patients at a higher risk of falls if it requires that they push the limits of their capacity. Alternatively, it is possible that, through repeated stepping practice, post- stroke individuals have reinforced a suboptimal pattern due to insufficient experience with a more optimal, symmetric pattern. Addressing these issues requires a thorough understanding of the processes by which stroke survivors optimize their walking pattern. Here, these issues are addressed using behavioral approaches to quantify tradeoffs between asymmetry and measures of walking ability such as stability and economy, and computational methods to identify the causal relationships linking these variables. Ultimately, the knowledge derived from this work will provide a mechanistic understanding of how the damaged brain optimizes movement, and may also inform the way in which clinicians develop personalized rehabilitation objectives for stroke survivors. Our findings may also inform cost/benefit analyses of walking in other patient populations known to have asymmetric walking patterns such as amputees or individuals with Parkinson's disease. Furthermore, providing a more mechanistic rationale for gait rehabilitation interventions could improve the efficiency of physical therapy, reduce health care costs, and ultimately help to better reintegrate individuals with neuromotor impairments into society by maximizing their mobility.